O07 Serious infection risk with systemic treatments for psoriasis: a cohort study from the British Association of Dermatologists Biologics and Immunomodulators Register (BADBIR)

Abstract Systemic treatments for psoriasis could increase the risk of serious infection due to their immunomodulatory properties. Our aim was to assess the risk of serious infection associated with biologics and systemic nonbiologics for psoriasis management. Adults with moderate-to-severe psoriasis from the British Association of Dermatologists Biologics and Immunomodulators Register (BADBIR) were eligible for this longitudinal cohort study. Inclusion criteria were adults who received at least one of the biologics, apremilast or conventional nonbiologics (acitretin, ciclosporin and methotrexate) for ≥ 6 months. All biologics licensed for psoriasis were analysed except for infliximab, which had higher prescription criteria. Conventional nonbiologics were grouped together. Patients were followed up from the initiation of treatment until the first serious infection episode, treatment switch or withdrawal, death or last available follow-up date. Infections resulting in hospitalization, administration of intravenous antibiotics or death were classified as serious. Incidence rates (IRs) of serious infection per 1000 person-years and corresponding confidence intervals were calculated using R version 4.3.2 (R Foundation, Vienna, Austria). In total 17 515 patients undergoing 32 759 treatment episodes were analysed. Patients were predominantly male in terms of treatment episodes (n = 18 633, 56.9%). The mean age was 46 years (SD 13.7), the mean body mass index 31.1 kg m−2 (SD 9.88) and the median treatment duration 2.69 years [interquartile range (IQR) 1.29–5.25]. The groups were similar except for age, treatment duration and concomitant psoriatic arthritis. The certolizumab group had a low mean age of 37.4 years (SD 10.3), the ustekinumab group had a significantly longer median treatment duration of over 4 years (IQR 1.83–6.73), and more patients in the ixekizumab (42.6%) and certolizumab (40.6%) groups had concomitant psoriatic arthritis. Overall, 1817 first serious infection episodes were recorded. The IR per 1000 person-years of the overall population was 14.6 [95% confidence interval (CI) 13.9–15.2]. The IR was lowest for apremilast (IR 7.52, 95% CI 2.44–17.6) and highest for certolizumab (IR 23.1, 95% CI 13.5–37.1). IRs (95% CIs) of other tumour necrosis factor-α inhibitors were 15.7 (14.5–17.1) for adalimumab and 16.7 (13.8–20.0) for etanercept. For interleukin-17 inhibitors the IRs (95% CIs) were 18.4 (15.9–21.2) for secukinumab, 7.63 (0.92–27.6) for bimekizumab, 14.5 (7.73–24.8) for brodalumab and 18.5 (14.0–24.0) for ixekizumab. For interleukin-23 inhibitors the IRs (95% CIs) were 13.5 (9.97–17.8) for guselkumab, 13.8 (9.10–20.1) for risankizumab and 17.2 (7.43–33.9) for tildrakizumab. The interleukin-12/23 inhibitor ustekinumab had an IR of 14.2 (95% CI 12.9–15.7) and for conventional nonbiologics the IR was 11.5 (95% CI 10.4–12.8). In this analysis, we observed a low incidence of serious infections across treatment groups. The Cox proportional hazards analysis with propensity scores and inverse probability treatment weighting to account for differences in potential confounders would provide further insights into the relative risk of serious infection.